Welcome back to the #33 edition of The New Defense Post!
In this edition, we’ll cover:
Spotlights: ICEYE secures a €300M revolving credit facility as sovereign space intelligence hits profitability; Helsing and OHB launch the KIRK joint venture, Europe's first named orbital targeting architecture; Stratolaunch discloses a Mach 5+ hypersonic test for the US Missile Defense Agency, and why reusable test infrastructure is the gap Europe keeps ignoring.
Fundraising News of the Week: ICEYE's €300M RCF anchors the table; Anduril closes a $5B Series H at $61B valuation just outside the window but too significant to skip.
Bonus Section: Shield AI and Taiwan's Thunder Tiger sign an MOU to run Hivemind on autonomous surface vessels, and what that tells founders about autonomous maritime tech for GPS-degraded coastal environments.
Spotlights
1. ICEYE Closes a €300M Credit Facility — and Proves the European Space Intelligence Model Works
Photo Credit: ICEYE
Finnish SAR satellite operator ICEYE announced on May 21 that it has originated a €300 million three-year revolving credit facility, backed by a seven-bank syndicate led by Citi and Danske Bank. (ICEYE) (EU-Startups)
The RCF isn't a fundraiser in the traditional sense. ICEYE will use it primarily to issue guarantees for customer contracts, which is how large-scale government procurement actually works: a ministry signs a frame agreement, and the contractor needs to post financial guarantees against delivery milestones before a single satellite gets built.
The fact that a seven-bank commercial syndicate is now backstopping ICEYE's government contract pipeline tells us something straightforward: ICEYE has moved from spacetech startup to creditworthy institutional supplier.
Technical note: SAR, or Synthetic Aperture Radar, produces high-resolution imagery by processing radar returns across a wide synthetic antenna aperture as the satellite moves along its orbit. Unlike optical sensors, SAR works in all weather and at night, which is the core reason defense customers value it over conventional Earth observation. ICEYE's SAR constellation is designed for short revisit times: the ability to image the same location multiple times per day.
At 25 cm resolution on its latest six satellites (launched March 2026), that's enough to track vehicle movements and infrastructure changes in near-real time. Deploying this as a sovereign system, rather than accessing commercial imagery through a US-controlled provider, is the strategic shift European militaries are now willing to pay for.
📰 Our Take: The Polish MikroSAR delivery in under 12 months is the number founders should be writing down. Most European defense procurement takes five to ten years from contract to delivery.
ICEYE did it in twelve months, on a sovereign system, for an MoD customer. The commercial RCF is proof the model scales: once you have a contract pipeline of this size, debt markets open up and equity dilution becomes optional.
Avoid dilutive financing when you can; debt is the best way forward when you start to have predictable revenues.
2. Helsing and OHB Launch KIRK, Europe's First Named Orbital Targeting Architecture
Photo Credit: OHB, Helsing, Hensoldt, Kongsberg
Munich-based defense AI company Helsing and Bremen satellite prime OHB announced on May 19 the formation of KIRK, a joint venture to build a space-based tactical surveillance, reconnaissance, and targeting system.
KIRK, short for Künstliche Intelligenz und Raumfahrt-Kompetenz (Artificial Intelligence and Space Competence), also formalises joint leadership of a wider consortium that includes Kongsberg Defence & Aerospace and HENSOLDT. OHB is the fourth partner to join, adding satellite manufacturing to a consortium that previously brought together AI software, sensors, and communications infrastructure. (Helsing) (AeroTime)
Technical note: Space-based targeting requires closing a sensor-to-shooter loop from orbit. A satellite passes over a target area for only a few minutes per orbit, so the window to collect, process, and transmit actionable data is narrow. Onboard AI processing addresses this by performing target recognition directly on the satellite before downlink, rather than sending raw imagery to a ground station for analysis.
HENSOLDT's space-qualified sensors provide all-weather persistent surveillance; Kongsberg's KSAT ground station network provides global downlink coverage; Helsing closes the intelligence loop with automated target recognition. The "software-defined" architecture means new sensor modes or target classification algorithms can be pushed to the constellation via software update, which is how you keep a space system relevant across a decade of threat evolution without launching new hardware.
📰 Our Take: Helsing contributes combat-proven AI; OHB contributes satellite manufacturing; Kongsberg contributes communications and ground infrastructure; HENSOLDT contributes sensors. Those are four complementary and specific capabilities, but it’s also easier to win as a consortium.
3. Stratolaunch Discloses a Mach 5+ Hypersonic Test for the MDA — and Reusability Is the Point
Photo Credit: Stratolaunch
On May 21, Stratolaunch announced that it had completed a successful hypersonic flight test for the US Missile Defense Agency on March 6, using its Talon-A3 reusable test vehicle and Spirit of Mojave, a modified Boeing 747-400 air-launch platform.
The vehicle, designated FEX-04, achieved Mach 5+ and was the company's third known hypersonic flight. Talon-A3 is powered by Ursa Major's Hadley liquid-fuelled engine. The vehicle was recovered after the flight, maintaining the reusable model that defines Stratolaunch's commercial proposition. (Defense Daily) (FlightGlobal)
Technical note: Reusable hypersonic test vehicles solve a specific problem in weapons development: the test environment itself is destructive and expensive. A single expendable hypersonic test vehicle costs tens of millions of dollars and is gone after one flight.
Stratolaunch's Talon-A is air-launched from a 747 at high altitude, accelerates to Mach 5+ under rocket power, gathers aerothermal and aerodynamic data in the hypersonic regime (above 20 km, above Mach 5), then glides back to a runway recovery. The reusable model reduces per-test cost significantly and, more importantly, increases test cadence.
📰 Our Take: Europe doesn't have a Stratolaunch equivalent. That isn't a criticism of any specific company — it's a structural gap that we caused ourselves.
Hypersonics are the future of long-range deep strikes. They are very difficult to counter, complex to build, and a game changer if built at scale.
Other News
Anduril Raises $5B Series H at $61B Valuation, Doubling Revenue to $2.2B in 2025 (Washington Technology)
UK Awards ~£1B Contract for 72 RCH 155 Remote-Controlled Howitzers to ARTEC, First Deliveries 2028 (Defence Industry Europe)
Poland Receives First Three F-35A Husarz Fighters at 32nd Tactical Air Base in Łask (Defence Industry Europe)
RTX BBN Technologies Demonstrates Self-Healing Communications System for Combat Air Support in Jammed Networks (Defence Industry Europe)
Fundraising News
Amount | Name | Round | Category |
|---|---|---|---|
€300M | SAR Satellite Intelligence / Space Defence | ||
$5B | Autonomous Defense Systems | ||
~$24.7M* | Reusable Hypersonic Test Platform |
Pilotix is a European drone technology company building end-to-end solutions, from high-performance drones and flight electronics to advanced software systems. With in-house production and high-precision assembly capabilities, Pilotix delivers reliable, scalable platforms for both civilian and defense applications, including surveillance, industrial operations, and specialised missions.
Use code HOBBYDRONEF1 for a discount.
Bonus Section — Hivemind Goes to Sea : What Shield AI's Taiwan Partnership Tells Founders About Autonomous Maritime Tech
Photo Credit: Taiwan News
On May 12-13, Shield AI and Taiwan's Thunder Tiger Corporation signed an MOU to integrate the Hivemind autonomy software into Thunder Tiger's unmanned surface vessel portfolio, starting with the SeaShark platform. The first milestone is a live AI-piloted maritime demonstration planned for summer 2026. A human won't be in the loop for navigation or mission execution during the demo.
Hivemind is Shield AI's core autonomy platform, the same software that flies F-16s without GPS. The key capability it brings to a surface vessel is the ability to operate in GPS-denied and communications-degraded environments without falling back on a human operator for every decision. In a contested maritime environment, where an adversary is actively jamming GPS and communications links, a USV dependent on constant human control is operationally equivalent to a boat with no skipper.
Technical note: Hivemind's approach to GPS-denied navigation on an aerial platform uses SLAM (Simultaneous Localisation and Mapping) fused with onboard sensor data. Translating this to a maritime surface vessel introduces a different sensor problem.
At sea, the environment is dynamic (wave state, vessel traffic, variable shoreline), the electromagnetic environment is dense (radar returns from ships, clutter from coastlines), and the relevant hazards are often below the surface line-of-sight. A maritime Hivemind integration needs to fuse radar, EO/IR, AIS transponder data, and acoustic sensors into a single situational picture.
Thunder Tiger's SeaShark family is already production-ready. The SeaShark 600 Stealth variant carries up to 600 kg of payload in a low-radar-cross-section hull, with a documented integration-to-demonstration timeline from Shield AI of under six weeks on previous platform types. Thunder Tiger operates in a hardware-native space that European equivalents are still working to match: the company has production capacity, a track record of delivering military-grade unmanned platforms, and now an AI autonomy partner with proven multi-domain credentials.
The partnership structure is the thing worth reverse-engineering. Shield AI provides the autonomy stack; Thunder Tiger provides the hull, the payload integration know-how, and the customer relationships with Taiwan's MND. Neither company tried to own the full stack.
GPS-denied, comms-degraded autonomous maritime operation is the requirement. For founders in sensor fusion, marine robotics, or edge-inference AI: the Speartooth LUUV delivered to the US Navy by Australia's C2 Robotics this month, the US Navy's 47-MUSV plan published this week, and this MOU are three separate signals pointing at the same gap.
The autonomous maritime layer is being built now. The software architectures going in first will be very hard to displace.
Love these insights? Forward this newsletter to a friend or two. They can subscribe to our website.